commit 17a17bf50612e6048a9975450cf1bd30f93815b5 upstream.
The mmc core uses a PM notifier to temporarily during system suspend, turn
off the card detection mechanism for removal/insertion of (e)MMC/SD/SDIO
cards. Additionally, the notifier may be used to remove an SDIO card
entirely, if a corresponding SDIO functional driver don't have the system
suspend/resume callbacks assigned. This behaviour has been around for a
very long time.
However, a recent bug report tells us there are problems with this
approach. More precisely, when receiving the PM_SUSPEND_PREPARE
notification, we may end up hanging on I/O to be completed, thus also
preventing the system from getting suspended.
In the end what happens, is that the cancel_delayed_work_sync() in
mmc_pm_notify() ends up waiting for mmc_rescan() to complete - and since
mmc_rescan() wants to claim the host, it needs to wait for the I/O to be
completed first.
Typically, this problem is triggered in Android, if there is ongoing I/O
while the user decides to suspend, resume and then suspend the system
again. This due to that after the resume, an mmc_rescan() work gets punted
to the workqueue, which job is to verify that the card remains inserted
after the system has resumed.
To fix this problem, userspace needs to become frozen to suspend the I/O,
prior to turning off the card detection mechanism. Therefore, let's drop
the PM notifiers for mmc subsystem altogether and rely on the card
detection to be turned off/on as a part of the system_freezable_wq, that we
are already using.
Moreover, to allow and SDIO card to be removed during system suspend, let's
manage this from a ->prepare() callback, assigned at the mmc_host_class
level. In this way, we can use the parent device (the mmc_host_class
device), to remove the card device that is the child, in the
device_prepare() phase.
Reported-by: Kiwoong Kim <kwmad.kim@samsung.com>
Cc: stable@vger.kernel.org # v4.5+
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20210310152900.149380-1-ulf.hansson@linaro.org
Reviewed-by: Kiwoong Kim <kwmad.kim@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 147186f531ae49c18b7a9091a2c40e83b3d95649 upstream.
A CMD11 is sent to the SD/SDIO card to start the voltage switch procedure
into 1.8V I/O. According to the SD spec a power cycle is needed of the
card, if it turns out that the CMD11 fails. Let's fix this, to allow a
retry of the initialization without the voltage switch, to succeed.
Note that, whether it makes sense to also retry with the voltage switch
after the power cycle is a bit more difficult to know. At this point, we
treat it like the CMD11 isn't supported and therefore we skip it when
retrying.
Signed-off-by: DooHyun Hwang <dh0421.hwang@samsung.com>
Link: https://lore.kernel.org/r/20210210045936.7809-1-dh0421.hwang@samsung.com
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 97fce126e279690105ee15be652b465fd96f9997 upstream.
In command queueing mode, the cache isn't flushed via the mmc_flush_cache()
function, but instead by issuing a CMDQ_TASK_MGMT (CMD48) with a
FLUSH_CACHE opcode. In this path, we need to check if cache has been
enabled, before deciding to flush the cache, along the lines of what's
being done in mmc_flush_cache().
To fix this problem, let's add a new bus ops callback ->cache_enabled() and
implement it for the mmc bus type. In this way, the mmc block device driver
can call it to know whether cache flushing should be done.
Fixes: 1e8e55b670 (mmc: block: Add CQE support)
Cc: stable@vger.kernel.org
Reported-by: Brendan Peter <bpeter@lytx.com>
Signed-off-by: Avri Altman <avri.altman@wdc.com>
Tested-by: Brendan Peter <bpeter@lytx.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20210425060207.2591-2-avri.altman@wdc.com
Link: https://lore.kernel.org/r/20210425060207.2591-3-avri.altman@wdc.com
[Ulf: Squashed the two patches and made some minor updates]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 21e35e898aa9ef7781632959db8613a5380f2eae upstream.
For data read commands, SDHC may initiate data transfers even before it
completely process the command response. In case command itself fails,
driver un-maps the memory associated with data transfer but this memory
can still be accessed by SDHC for the already initiated data transfer.
This scenario can lead to un-mapped memory access error.
To avoid this scenario, reset SDHC (when command fails) prior to
un-mapping memory. Resetting SDHC ensures that all in-flight data
transfers are either aborted or completed. So we don't run into this
scenario.
Swap the reset, un-map steps sequence in sdhci_request_done().
Suggested-by: Veerabhadrarao Badiganti <vbadigan@codeaurora.org>
Signed-off-by: Pradeep P V K <pragalla@codeaurora.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/1614760331-43499-1-git-send-email-pragalla@qti.qualcomm.com
Cc: stable@vger.kernel.org # v4.9+
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 774514bf977377c9137640a0310bd64eed0f7323 upstream.
An issue has been observed on STM32MP157C-EV1 board, with an erase command
with secure erase argument, ending up waiting for ~4 hours before timeout.
The requested busy timeout from the mmc core ends up with 14784000ms (~4
hours), but the supported host->max_busy_timeout is 86767ms, which leads to
that the core switch to use an R1 response in favor of the R1B and polls
for busy with the host->card_busy() ops. In this case the polling doesn't
work as expected, as we never detects that the card stops signaling busy,
which leads to the following message:
mmc1: Card stuck being busy! __mmc_poll_for_busy
The problem boils done to that the stm32 variants can't use R1 responses in
favor of R1B responses, as it leads to an internal state machine in the
controller to get stuck. To continue to process requests, it would need to
be reset.
To fix this problem, let's set MMC_CAP_NEED_RSP_BUSY for the stm32 variant,
which prevent the mmc core from switching to R1 responses. Additionally,
let's cap the cmd->busy_timeout to the host->max_busy_timeout, thus rely on
86767ms to be sufficient (~66 seconds was need for this test case).
Fixes: 94fe2580a2 ("mmc: core: Enable erase/discard/trim support for all mmc hosts")
Signed-off-by: Yann Gautier <yann.gautier@foss.st.com>
Link: https://lore.kernel.org/r/20210225145454.12780-1-yann.gautier@foss.st.com
Cc: stable@vger.kernel.org
[Ulf: Simplified the code and extended the commit message]
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit c5b1c6dc13daec60405ecd31eaa5379a9f798fa8 ]
The device_* calls were added a few years ago to abstract
DT/ACPI/fwnode firmware interfaces. Lets convert the two
sdhci caps fields to use the generic calls rather than the OF
specific ones. This has the side effect of allowing
ACPI based devices to quirk themselves when the caps field
is broken.
Signed-off-by: Jeremy Linton <jeremy.linton@arm.com>
Link: https://lore.kernel.org/r/20201120233831.447365-1-jeremy.linton@arm.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4f9833d3ec8da34861cd0680b00c73e653877eb9 ]
The RPi4 has an Arasan controller it carries over from the RPi3 and a newer
eMMC2 controller. Because of a couple of quirks, it seems wiser to bind
these controllers to the same driver that DT is using on this platform
rather than the generic sdhci_acpi driver with PNP0D40.
So, BCM2847 describes the older Arasan and BRCME88C describes the newer
eMMC2. The older Arasan is reusing an existing ACPI _HID used by other OSes
booting these tables on the RPi.
With this change, Linux is capable of utilizing the SD card slot, and the
Wi-Fi when booted with UEFI+ACPI on the RPi4.
Signed-off-by: Jeremy Linton <jeremy.linton@arm.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210120000406.1843400-2-jeremy.linton@arm.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 0354ca6edd464a2cf332f390581977b8699ed081 ]
when get request SW timeout, if CMD/DAT xfer done irq coming right now,
then there is race between the msdc_request_timeout work and irq handler,
and the host->cmd and host->data may set to NULL in irq handler. also,
current flow ensure that only one path can go to msdc_request_done(), so
no need check the return value of cancel_delayed_work().
Signed-off-by: Chaotian Jing <chaotian.jing@mediatek.com>
Link: https://lore.kernel.org/r/20201218071611.12276-1-chaotian.jing@mediatek.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit d7fb9c24209556478e65211d7a1f056f2d43cceb upstream.
The implementation of sdhci_pltfm_suspend() is only available when
CONFIG_PM_SLEEP is set, which triggers a linking error:
"undefined symbol: sdhci_pltfm_suspend" when building sdhci-brcmstb.c.
Fix this by implementing the missing stubs when CONFIG_PM_SLEEP is unset.
Reported-by: Arnd Bergmann <arnd@arndb.de>
Suggested-by: Florian Fainelli <f.fainelli@gmail.com>
Fixes: 5b191dcba719 ("mmc: sdhci-brcmstb: Fix mmc timeout errors on S5 suspend")
Cc: stable@vger.kernel.org
Tested-By: Nicolas Schichan <nschichan@freeebox.fr>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 5b191dcba719319148eeecf6ed409949fac55b39 upstream.
Commit e7b5d63a82 ("mmc: sdhci-brcmstb: Add shutdown callback")
that added a shutdown callback to the diver, is causing "mmc timeout"
errors on S5 suspend. The problem was that the "remove" was queuing
additional MMC commands after the "shutdown" and these caused
timeouts as the MMC queues were cleaned up for "remove". The
shutdown callback will be changed to calling sdhci-pltfm_suspend
which should get better power savings because the clocks will be
shutdown.
Fixes: e7b5d63a82 ("mmc: sdhci-brcmstb: Add shutdown callback")
Signed-off-by: Al Cooper <alcooperx@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20210107221509.6597-1-alcooperx@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1a3ed0dc3594d99ff341ec63865a40519ea24b8d upstream.
Automatic Clock Gating is a feature used for the power consumption
optimisation. It turned out that during early init phase it may prevent the
stable voltage switch to 1.8V - due to that on some platforms an endless
printout in dmesg can be observed: "mmc1: 1.8V regulator output did not
became stable" Fix the problem by disabling the ACG at very beginning of
the sdhci_init and let that be enabled later.
Fixes: 3a3748dba8 ("mmc: sdhci-xenon: Add Marvell Xenon SDHC core functionality")
Signed-off-by: Alex Leibovich <alexl@marvell.com>
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
Cc: stable@vger.kernel.org
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20201211141656.24915-1-mw@semihalf.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ca1219c0a7432272324660fc9f61a9940f90c50b upstream.
Commit a44f7cb937 ("mmc: core: use mrq->sbc when sending CMD23 for
RPMB") began to use ACMD23 for RPMB if the host supports ACMD23. In
RPMB ACM23 case, we need to set bit 31 to CMD23 argument, otherwise
RPMB write operation will return general fail.
However, no matter V4 is enabled or not, the dwcmshc's ARGUMENT2
register is 32-bit block count register which doesn't support stuff
bits of CMD23 argument. So let's handle this specific ACMD23 case.
From another side, this patch also prepare for future v4 enabling
for dwcmshc, because from the 4.10 spec, the ARGUMENT2 register is
redefined as 32bit block count which doesn't support stuff bits of
CMD23 argument.
Fixes: a44f7cb937 ("mmc: core: use mrq->sbc when sending CMD23 for RPMB")
Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20201229161625.38255233@xhacker.debian
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The #ifdef check for the suspend/resume functions is wrong:
drivers/mmc/host/mtk-sd.c:2765:12: error: unused function 'msdc_suspend' [-Werror,-Wunused-function]
static int msdc_suspend(struct device *dev)
drivers/mmc/host/mtk-sd.c:2779:12: error: unused function 'msdc_resume' [-Werror,-Wunused-function]
static int msdc_resume(struct device *dev)
Remove the #ifdef and mark all four as __maybe_unused to aovid the
problem.
Fixes: c0a2074ac5 ("mmc: mediatek: Fix system suspend/resume support for CQHCI")
Cc: stable@vger.kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20201203222922.1067522-1-arnd@kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The CMD13 polling is needed for commands with R1B responses. In commit
a0d4c7eb71 ("mmc: block: Add CMD13 polling for MMC IOCTLS with R1B
response"), the intent was to introduce this for requests targeted to the
RPMB partition. However, the condition to trigger the polling loop became
wrong, leading to unnecessary polling. Let's fix the condition to avoid
this.
Fixes: a0d4c7eb71 ("mmc: block: Add CMD13 polling for MMC IOCTLS with R1B response")
Cc: stable@vger.kernel.org
Reported-by: Zhan Liu <zliua@micron.com>
Signed-off-by: Zhan Liu <zliua@micron.com>
Signed-off-by: Bean Huo <beanhuo@micron.com>
Link: https://lore.kernel.org/r/20201202202320.22165-1-huobean@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
The commit 16ada730a7 ("mmc: sdhci-of-arasan: Modify clock operations
handling") introduced support for platform specific clock operations.
Around the same point in time the commit 36c6aadaae ("mmc:
sdhci-of-arasan: Add support for Intel Keem Bay") was also merged.
Unfortunate it was not really tested on top of the previously mentioned
commit, which causes clock registration failures for Keem Bay SOC devices.
Let's fix this, by properly declaring the clock operation for Keem Bay SOC
devices.
Fixes: 36c6aadaae ("mmc: sdhci-of-arasan: Add support for Intel Keem Bay")
Signed-off-by: Muhammad Husaini Zulkifli <muhammad.husaini.zulkifli@intel.com>
Reviewed-by: Adrian Hunter <adrian.hunter@intel.com>
Link: https://lore.kernel.org/r/20201118120120.24908-2-muhammad.husaini.zulkifli@intel.com
Cc: stable@vger.kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
A UHS setting of SDR25 can give better results for High Speed mode.
This is because there is no setting corresponding to high speed. Currently
SDHCI sets no value, which means zero which is also the setting for SDR12.
There was an attempt to change this in sdhci.c but it caused problems for
some drivers, so it was reverted and the change was made to sdhci-brcmstb
in commit 2fefc7c5f7 ("mmc: sdhci-brcmstb: Fix incorrect switch to HS
mode"). Several other drivers also do this.
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: stable@vger.kernel.org # v5.4+
Link: https://lore.kernel.org/r/20201112133656.20317-1-adrian.hunter@intel.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
